Artificial intelligence shaping ESG risk governance in banking: Evidence from a systematic literature review

Abstract

This paper examines how risk governance architecture shapes interactions between artificial intelligence (AI) and environmental, social, and govern­ance (ESG)-oriented sustainability policies in the banking industry. Most current research treats AI as a technological capability that directly affects ESG performance, yet little is known about the governance systems that produce these outcomes. Using PRISMA-guided SLR procedures, we selected 20 studies from 248 initial records identified in the Scopus and Web of Science databases that met the inclusion criteria and conducted a thematic synthesis. The results show that AI is primarily used in ESG disclosure and reporting, credit risk assessment, climate risk analytics, sustainable finance, and responsible AI governance. The literature remains dispersed across theo­retical stances, including the resource-based view, stakeholder theory, institutional theory, legitimacy theory, and AI governance literature. Previ­ous research has largely ignored the governance mechanisms that enable successful implementation, focusing instead on the direct implications of AI adoption for ESG-related outcomes. The study proposes an AI–ESG risk governance integrative framework to address this gap. This framework places risk governance architecture at the center of the relationship among AI capabilities, institutional pressures, stakeholder expectations, and ESG-oriented sustainability outcomes. The approach views AI as a strategic capacity integrated into enterprise-wide risk governance systems rather than merely a technical or compliance tool. The results indicate that strong governance arrangements, such as model governance, accountability frame­works, board supervision, and alignment with organizational risk appetite, are necessary for successfully deploying AI-enabled ESG. By offering an integrative theoretical framework and practical insights for banking organi­zations seeking to improve sustainability performance and long-term resilience through responsible AI use, this study contributes to the growing body of AI-ESG literature.